MiniMax-M1 is a reasoning-focused model with a massive context window of up to one million tokens and a "thinking" budget of up to 80,000 tokens. The model uses an especially efficient reinforcement ...

Luxembourg’s armed forces are also on board. The defense ministry has signed its own contract with Mistral AI, aiming to integrate AI tools into military operations. The government says all AI ...

Google launches Audio Overviews in search results Google is rolling out a new feature called Audio Overviews in its Search Labs. Powered by the Gemini language model, Audio Overviews automatically ...

TikTok has introduced three new AI tools—"Image to Video," "Text to Video," and "Showcase Products"—that enable advertisers to create video content in a more cost-effective way. These tools ...

OpenAI has significantly updated ChatGPT's search feature: it now handles longer contexts, better follows instructions, answers complex questions with several parallel searches, and allows users to ...

Fundamental disagreements over AI's future LeCun's remarks highlight a much deeper debate about the direction of AI research. Companies like Anthropic and OpenAI are racing to commercialize ever more ...

OpenAI has lowered the price of its o3 language model by 80 percent, CEO Sam Altman said. The new cost is $2 per million input tokens and $8 per million output tokens. The move follows Google’s Gemini ...

Generative AI is quietly changing the UK's film, TV, and gaming industries, according to a new report from the British Film Institute (BFI). The study finds AI is already used behind the scenes to ...

According to OpenAI, the revamped "Advanced Voice Mode" now delivers smoother, more emotionally nuanced speech, with improvements to intonation, pauses, and the ability to convey empathy or sarcasm in ...

Large Reasoning Models (LRMs) such as Claude 3.7 Sonnet Thinking, Deepseek-R1, and OpenAI's o3 are often described as a step toward more general artificial intelligence. Techniques like ...

ByteDance unveiled Seedance 1.0, a new AI model for video generation. According to the company's tests, Seedance 1.0 outperforms established systems such as Google's Veo, Kuaishou's Kling, and ...

This puts Magistral roughly on par with the first Deepseek-R1 model from January, but still well behind the recently updated R1-0528 as well as OpenAI's o3 and Google's latest Gemini Pro 2.5. It seems ...